bio |
professional |
professional |
personal |
personal |
pretend |
pretend |
|
work |
papers |
papers |
psychsim |
psychsim |
plan recog |
plan recog |
|
play |
pynadath.net |
pynadath.net |
poker |
poker |
p evil AI |
p evil AI |
|
|
|
I am a Computer Scientist at the USC
Institute for Creative Technologies.
But I have a confession to make: I don’t like computers. Now, I’m
not one of those elitists who look down on computers just because of
their immense stupidity. After all, if my computer had any natural
intelligence, I’d be out of a job. And he has proven to be
occasionally useful at performing menial tasks, like
database
queries and Web hosting. But I find myself constantly infuriated by my computer’s complete failure as a friend, due to his deeply entrenched inability to grasp the fine points of social intercourse.
I once asked my computer, “How many people would get upset if
I—” “42,” he interrupted. Incredulous, I asked, “How’d you come up with that answer? You didn’t even let me finish!” He shrugged, “Does it matter? Just trust me.” I warily asked, “How do I know that you’re right?” He sighed, “How would you know if I’m wrong?” His cynicism and hubris get on my nerves after a while.
But there are also times when he becomes inconveniently modest about his capabilities. One day, I asked for some help getting out of a jam with a man-eating tiger. Instead, I got a lot of foot dragging. He first tells me, “There’re over 6×1014 possible policies.” I said, “What?!” He insisted, “It would take me almost 20 million years to find the best answer.” I started to get annoyed, “Just make a guess!” He glanced uneasily around and said, “Even with the best policy, my expected utility would still be negative.” Exasperated, I broke down, “C’mon, we’re a team!” But he was having none of it and left me to fend for myself.
While recovering in the hospital, I had time to reflect on his behavior, and I began to suspect a sinister purpose. Call me paranoid, but I recognize a plot when I see
one. You may think it’s unlikely that my computer is out to get me, but it’s enough of a possibility for me to be on my guard. You would be wise to be similarly vigilant.
Email:
pynadath@ict.usc.edu
|
|
|
|

I grew up in Johnstown, NY, a small upstate town which has a population of a little over 9,000, and which is most famous for not being the site of the semi-famous Johnstown Flood (which took place in Johnstown, PA). However, it does have its own exit on the NY State Thruway, making leaving that much easier when it was time for college. I graduated from MIT in 1992 with S.B.’s in electrical engineering and computer science. I then worked at the Tufts Medical School on computer interfaces using eye tracking for disabled users without any other modes of input. After a year, the grant moved to a different institution and forgot to take us researchers with it, so I moved on to graduate school at the AI Lab of the University of
Michigan instead. There, I worked with Michael Wellman on a probabilistic framework for plan recognition, with a digression into stochastic grammars.
I defended my thesis (Probabilistic Grammars for Plan Recognition) in January, 1999, and began work as a research scientist at ISI in December, 1998. Since time travel was not a viable option at the time, my first two months here required an
overlap of the two efforts, a challenging experience to say the least (Public Service Announcement: Kids, stay in school). I’m currently working at ICT on social simulation with Stacy Marsella. I previously worked on the TEAMCORE project with Milind Tambe.
Since moving to LA, I feel like I spend most of my time trying to figure out ways to survive without a driver’s license. I’ve held out so far, but my resistance is cracking. My more enjoyable pursuits include playing the flute less frequently than I should, playing any sport I can fit into my schedule, rooting for Philadelphia sports teams, watching movies made with under $50 million, trying to become as well-read as my mother, and trying not to spend all of my money at used CD stores.
In an effort to boost my relatives’ Web presence, I include the following links:
|
|
|
|
[WARNING: I wrote the following “biography” back in
1991, so it is a bit dated. For instance, we now know that anyone who
plays the Riddler will never, ever be nominated for an Oscar for
it, or any other role.]
Love is not as relevant as money, but it is
more relevant than life.
This profound statement from his
essay, “Daze of My Youth,” epitomizes David Pynadath’s uncanny ability to
combine ironic observations of today’s society with idealistic hopes
for the future. He portrays an absurd world where it is impossible to find
meaning in anything, including his own essays. The simplicity with which he
states his observations belies the complex issues which we read into them.
In fact, because of their incisive vagueness, his essays incite more
discussion than any works other than the plays of Shakespeare and the
Far Side cartoons of Gary Larson.
Born in upstate New York, Pynadath was raised to be an engineer, complete
with a college education at MIT. “It took a lot of hard work, but it was
also a lot of fun,” he said of his school days, “though never at the same
time.” However, after taking a course in expository writing, he realized
the potential market for his talent and, upon graduation, decided to forego
graduate school and become a writer. He struggled at first, burdened by the
intense disapproval of his family and by the inability of any editor to
correctly pronounce his name.
His first job was writing lyrics for the pop group, New Kids on the Block,
an experience that drove home what would later become the guiding principle
of his writing: “Americans are morons.” He made a quick exit from the
music business and went to work for Madison Avenue. Advertising, he soon
realized, had little use for his talent, or anyone else’s talent for that
matter. But it did provide him with his first success as a writer when he
penned the Bush administration’s popular slogan, “Look at Dukakis in
Massachusetts, and just be grateful!” His ability was noticed by the
New York Times, and he replaced Russell Baker as the writer of the
Observer column. However, Baker’s name remained in the by-line for
a few years, while Pynadath stubbornly refused to change his name to Rajiv
Gandhi, as the Times wanted.
But the day when his true name was printed in the New York Times
finally came, and he became an instant celebrity. He published A Rose
Is a Rose Is a Buick (1995) and Hell Has Air-Conditioning (1996),
both of which became immediate best-sellers. True fame, which comes only to
those invited to Late Night with David Letterman, arrived soon after
Blank Expressions (1997), which sold more copies than any other book
except for the Bible and Chairman Mao’s Little Red Book (though as a
conversational topic at cocktail parties, it was more popular than the
former).
Soon to follow were an evening news program and a Saturday morning cartoon,
the latter based on his column. He also made appearances in Back
to the Future VI, as himself, and in Batman III, as the Riddler, a
role which won him the Academy Award for “Best Performance as a Comic Book
Villain in a Sequel.” However, in his Oscar acceptance speech, Pynadath
criticized the Academy’s voting procedure and failed to spend at least 15
minutes thanking people. Hollywood subsequently prohibited him from acting
in any more movie sequels, so future film appearances are unlikely.
Subsequent publication of How the West was Won (By Japan) (1998) and
Germany III and Other Comments on the State of the Universe (1999)
have only served to increase Pynadath’s public stature, and a Presidential
campaign seems likely. But his fame does not surprise him. “If Dan Quayle
could become Vice President without any qualifications whatsoever, then I
deserve at least the success which is mine today.” This confidence,
bordering on obnoxiousness, is his greatest strength. His sometimes awkward
phrasing forces the reader to concentrate on the essay, to make sure the
true meaning is conveyed. He blends the humor of E.B. White and the
technical ability of Woody Allen into a style that can only be his own.
When asked where his seemingly endless inspiration comes from, he turns
religious and says, “God only knows.” He offers the triweekly deadlines
of the Observer as a possible answer, “because when it’s five
minutes to deadline, the topic becomes the last thing on my mind. I just
write.” While this would blur the meaning of a lesser writer’s essay,
Pynadath’s occasional lack of focus increases his essay’s scope, making it
hit the reader closer to home, much like a horoscope. Fortunately, the
rigors of turning out word upon word upon word in inimitable fashion have
not worn down his love for writing. “On the contrary,” he said in a
recent interview, “I receive no greater pleasure than from watching people
praise a piece of work I put together in under an hour.” We can only
hope that more young writers will follow his example.
|
|
|
|
|
| Modeling Emotion as Byproduct of Theory of Mind Reasoning |
| Mei Si; Stacy Marsella; David Pynadath |
| Proceedings of the Intelligent Virtual Agents Conference, pp. 334-347, 2008 |
|
|
|
|
|
| Thespian: Modeling Socially Normative Behavior in a Decision-Theoretic Framework |
| Mei Si; Stacy Marsella; David V. Pynadath |
| Proceedings of the International Conference on Intelligent Virtual Agents, pp. 369-382, 2006 |
|
| Thespian: Using Multi-Agent Fitting to Craft Interactive Drama |
| Mei Si; Stacy C. Marsella; David V. Pynadath |
| Proceedings of the International Conference on Autonomous Agents and Multiagent Systems, pp. 21-28, 2005 |
|
|
|
|
|
|
|
|
|
|
|
| Team Coordination among Distributed Agents: Analyzing Key Teamwork Theories and Models |
| David V. Pynadath; Milind Tambe |
| Proceedings of the AAAI Spring Symposium on Intelligent Distributed and Embedded Systems, pp. 57-62, 2002 |
|
| Electric Elves: Adjustable Autonomy in Real-world Multi-Agent Environments |
| David V. Pynadath; Milind Tambe |
| Socially Intelligent Agents: Creating Relationships with Computers and Robots, pp. 101-108, 2002 |
|
|
|
|
|
|
| Revisiting Asimov's First Law: A Response to the Call to Arms |
| David V. Pynadath; Milind Tambe |
| Proceedings of the IJCAI Workshop on Autonomy, Delegation, and Control: Interacting with Autonomous Agents, 2001 |
|
| Revisiting Asimov's First Law: A Response to the Call to Arms |
| David V. Pynadath; Milind Tambe |
| Intelligent Agents VIII: Agent Theories, Architectures and Languages, pp. 307-320, 2001 |
|
| MDPs for Adjustable Autonomy in Real-World Multi-Agent Environments |
| David V. Pynadath; Paul Scerri; Milind Tambe |
| Proceedings of the AAAI Spring Symposium on Game Theoretic and Decision Theoretic Agents, pp. pp. 107-116, 2001 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| A Study of the Information Capacity of Human Eye Movement for Augmentative Communication |
| David V. Pynadath; Hong Z. Tan; David M. Horowitz |
| Proceedings of the Rehabilitation Engineering and Assistive Technology Society of North America Conference, pp. 96-98, 1993 |
|
|
Social Simulation
Under construction. In the meantime, here is some PsychSim documentation.
|
Plan, Activity, and Intent Recognition
Workshops
David V. Pynadath
Last modified: Fri Jul 27 17:07:48 PDT 2007
|
|
|
Poker
Under construction.
|
Evil AI @ The Movies
July, 2007
It’s summer blockbuster season, and that can mean only one thing: robots trying to destroy the world! In this year’s version of the age-old story, Transformers has Megatron and his fellow reconfigurable robots turning Earth into their personal battlefield. As AI researchers laboring within sight of the Hollywood sign, we cannot help but notice the continual stream of movies that depict the product of our field as an inevitable threat to mankind. We are dismayed at the potential negative image of AI that may form in the impressionable young minds who download these movies. It is imperative that we carefully monitor the image of our work presented in the mass media.
That is not why I am writing this series.
No, I have no idea what effect these movies are having on the potential AI researchers of tomorrow. For all I know, millions of kids are watching The Matrix and deciding that they want to study AI and get in on the ground floor of the machine revolution. My motivation in this endeavor is much more pragmatic: the search for inspiration. Science and science fiction have a long history of symbiosis, each feeding new ideas into the other. We have made great strides in transforming many fictional visions into scientific reality. For example, filmgoers in 1968 were impressed when HAL 9000 easily defeats his human opponent in chess, but today’s audiences take such an outcome for granted.
On the other hand, while reviewing candidate movies to form the basis of this series, I realized that Hollywood has identified one gaping hole in our field: AI systems that are capable of being truly evil. In this area, fiction far outstrips reality. Sure, we have made some progress (e.g., the Microsoft paper clip), but we should not be satisfied with such baby steps. Unfortunately, ambitious young researchers pursuing Hollywood’s vision of evil are hamstrung by the prejudices of our field. Papers, workshops, and programs on topics such as safe agents, human-in-the-loop systems, ethical AI, etc. all indicate a mindset predisposed to favor AI systems that serve humanity, rather than seek to enslave it. In this series, I seek to encourage the pursuit of evil AI by examining several movies that offer insight into the research challenges that must be overcome.
- Hal 9000 (from 2001: A Space Odyssey)
- Colossus (from Colossus: The Forbin Project))
- The False Maria (from Metropolis))
- Proteus IV (from Demon Seed)
David Pynadath
Last modified: Tue Sep 30 17:03:34 PDT 2008
|
Agent Teamwork
Under construction.
|
|